Four consecutive trials were undertaken to study lorazepam at each of three dosage levels and diphenhydramine when used in combination with iv metoclopramide and dexamethasone in patients receiving cisplatin at 120 mg/m2. The combination containing diphenhydramine had been the most effective treatment identified in prior trials. Earlier studies have found lorazepam to be a useful adjunct to other antiemetic agents. Sixty-five patients who had never received chemotherapy or antiemetics were directly observed in the hospital for 24 hours following cisplatin. Overall, 56% of the patients experienced no emesis and 78.5% of the patients had two or fewer vomiting episodes during the study period. No significant differences were noted in the number of patients who experienced no emesis or two or fewer episodes among the four trials. More sedation was seen with the lorazepam-containing regimens; other side effects were similar in type and severity. The trial using the highest dose of lorazepam (1.5 mg/m2) demonstrated significantly greater patient satisfaction (P = 0.039) and less anxiety during therapy (P = 0.02) when compared with the diphenhydramine combination. We conclude that combinations of iv metoclopramide plus dexamethasone with either diphenhydramine or lorazepam are well tolerated and effective in controlling cisplatin-induced emesis. The regimens containing lorazepam produced better subjective evaluations, and the combination using the highest lorazepam dose tested showed superior patient satisfaction and less anxiety during therapy.